T An Efficient Approach for Medical Image Segmentation Based on Truncated Skew Gaussian Mixture Model Using K - Means Algorithm

In this paper, we proposed a novel approach for medical image segmentation process based on Finite Truncated Skew Gaussian mixture model. This approach considers various issues like skewness and asymmetric distributions with a finite range. We have utilized the Expectation-Maximization (EM) algorithm in estimating the final model parameters and K-Means algorithm is utilized to estimate the number of mixture components along with the initial estimates for the EM algorithm. Segmentation of the image is performed based on the maximum likelihood estimation criteria. The performance of the approach is evaluated using segmentation quality metrics and the image quality metrics such as Average Distance, Maximum Difference, Image Fidelity, Mean Squared Error and Signal to Noise ratio.